Before I log a call, anyone out there know what the above error means?

I know it's probably something to do with the optimizer. And it's a "big
query" that's causing it.

Any ideas?

Thanks



Gareth



Cymraeg:-
Mae'r neges hon yn gyfrinachol nad chi yw'r derbynnydd y bwriedid y neges ar ei gyfer, byddwch mor garedig â rhoi gwybod
i'r anfonydd yn ddi-oed. Dylid ystyried un rhywd datganiadau neu sylwadau a wneir uchod yn rhai personol,ac nid o angen rhaid yn rhai o
eiddo Ymddiriedolaeth GIG Bro Morgannwg, nac unrhyw ran gyfansoddol ohoni na chorff cysylltiedig.

Cofiwch fod yn ymwybodol ei bod yn bosibl y bydd disgwyl i Ymddiriedolaeth GIG Bro Morgannwg roi cyhoeddusrwydd i gynnwys unrhyw ebost neu
ohebiaeth a dderbynnir, yn unol ag amodau'r Ddeddf Rhyddid Gwybodaeth 2000. I gael mwy o wybodaeth am Ryddid Gwybodaeth, cofiwch gyfeirio
at wefan Ymddiriedolaeth GIG Bro Morgannwg ar www.bromor-tr.wales.nhs.uk

English:-
This message is confidential. If you are not the intended recipient of the message then please notify the sender immediately.
Any of the statements or comments made above should be regarded as personal and not necessarily those of Bro Morgannwg NHS Trust, any
constituent part or connected body.

Please be aware that, under the terms of the Freedom of Information Act 2000, Bro Morgannwg NHS Trust may be required to make public the
content of any emails or correspondence received. For further information on Freedom of Information, please refer to the Bro Morgannwg NHS
Trust website at www.bromor-tr.wales.nhs.uk.

On Mar 6, 2008, at 8:29 AM, Gareth Williams wrote:


Strangely enough, Jim (Gramling) and I just found the cause of a
sporadic
OP0889 on Monday. We know what to fix to make the error go away, but
we're still researching the root cause (there are two bugs that have to
interact to cause our OP0889). This is in 9.x using the new
enumeration (greedy enumeration). We'll submit this fix to Ingres
Corp as soon as we figure out the other half of the problem.
Your OP0889 may or may not be the same thing.

Anyway, go ahead and call it in. The error means that the code
generation phase of the optimizer expected to reference an
"equivalence class" (a column, essentially) that should have
been passed up to the QEP node being compiled, but there had
been no code generated to make that eqclass available.
This is an internal error and always indicates a bug of
some sort.

Karl

Hi Gareth,



That error code is known to the knowledge base. Do you get E_OP04c0 as well? What version of ingres are you using? It might be worth your while to check the knowledg base with the error codes and then see if you have a fix in your patch for this problem.



Martin Bowes



From: info-ingres-bounces (AT) kettleriver...ting (DOT) com [mailto:info-ingres-bounces (AT) kettleriverconsulting (DOT) com] On Behalf Of Gareth Williams
Sent: 06 March 2008 13:30
To: info-ingres (AT) kettleriverconsulting (DOT) com
Subject: [Info-Ingres] E_OP0889 Eqc is not available at a CO node.



Before I log a call, anyone out there know what the above error means?

I know it's probably something to do with the optimizer. And it's a "big query" that's causing it.

Any ideas?

Thanks



Gareth



Cymraeg:-
Mae'r neges hon yn gyfrinachol.Os nad chi yw'r derbynnydd y bwriedid y neges ar ei gyfer, byddwch mor garedig â rhoi gwybod i'r anfonydd yn ddioed. Dylid ystyried un rhywd datganiadau neu sylwadau a wneir uchod yn rhai personol,ac nid o angen rhaid yn rhai o eiddo Ymddiriedolaeth GIG Bro Morgannwg, nac unrhyw ran gyfansoddol ohoni na chorff cysylltiedig.

Cofiwch fod yn ymwybodol ei bod yn bosibl y bydd disgwyl i Ymddiriedolaeth GIG Bro Morgannwg roi cyhoeddusrwydd i gynnwys unrhyw ebost neu o hebiaeth a dderbynnir, yn unol ag amodau'r Ddeddf Rhyddid Gwybodaeth 2000. I gael mwy o wybodaeth am Ryddid Gwybodaeth, cofiwch gyfeirio at wefan Ymddiriedolaeth GIG Bro Morgannwg ar www.bromor-tr.wales.nhs.uk <http://www.bromor-tr.wales.nhs.uk/>

English:-
This message is confidential. If you are not the intended recipient of the message then please notify the sender immediately. Any of the statements or comments made above should be regarded as personal and not necessarily those of Bro Morgannwg NHS Trust, any constituent part or connected body.


Please be aware that, under the terms of the Freedom of Information Act 2000, Bro Morgannwg NHS Trust may be required to make public the content of any emails or correspondence received. For further information on Freedom of Information, please refer to the Bro Morgannwg NHS Trust website at www.bromor-tr.wales.nhs.uk <http://www.bromor-tr.wales.nhs.uk/> .

Hi,
we hit this problem too on Ingres 2006 a64.lnx and ingres support gave
us patch 11873.

It helped for some time but now we have it back again. Usualy it's sql
command like update tableA from tableB where id in (select id from
tableA).

The best solution is to regenerate statistics or you can change a
table structure to heap and add an index instead of table structure.

Hi,
I've got the problem on a select / union statement. Nothing all that
comlpicated either:
SELECT Mobilenumber, Q1.ref
FROM
(SELECT ref, phonenum1 as MobileNumber, notes FROM aisdba.syphone WHERE
applic_id='ct' AND Phonetype1=4
UNION
SELECT ref, phonenum2 as MobileNumber, notes FROM aisdba.syphone WHERE
applic_id='ct' AND Phonetype2=4
UNION
SELECT ref, phonenum3 as MobileNumber, notes FROM aisdba.syphone WHERE
applic_id='ct' AND Phonetype3=4) as Q1, (SELECT account_ref FROM t6230)
as T1
WHERE q1.ref = T1.account_ref
An ideas anyone...


--
Craig

The earlier thread that you have exhumed was referring to a stale
version of Ingres. You are clearly using a much later version that
supports subqueries in the FROM clause. Maybe an old bug has regressed,
or maybe it's something new.

If you've hit a bug it's hard to know what might work around it. Just
in the interests of making changes to see what happens, there's no
reason to use a subquery on t6230--it's reducing clarity and gives the
parser/optimizer more things to think about (and possibly get wrong).
You don't seem to use the *notes* attribute for anything so that can go
too.

Shame you're stuck with a table like aisdba.syphone, with it's
pseudo-repeating "fields".

All that aside, you shouldn't be getting weird errors on that query, so
if you have a support subscription raise it with tech support.


--
rhann